The present invention relates to a system and method for protecting information data and, in particular, for checking and granting authorization to access information data.

There are known data protection methods, in particular methods for authorizing access to protected information and/or allowing the execution of computer programs (software) subject to obtaining a license to use them.

Known methods contemplate the use of a key not know to the user and obtainable by purchasing a license to use the software in question. This method has the drawback that this key can be fraudulently replicated, so that unauthorized users who learn about this key are able to use the software.

Other methods contemplate the use of a device, called a token, able to generate a password, generally in a numeric format, on the basis of an algorithm unknown to the user. Typically, in a two-factor authentication scheme, this password assumes a value dependent on parameters that vary with time (e.g., time, date, etc.), plus a known part, chosen by the user (known as the Personal Identification Number or, more simply, the PIN) so as not to be easily replicable. This method has the drawback that everyone knows the time and the PIN is usually chosen by a user so that it is easy to remember (in practice, the PIN is often chosen as the user's date of birth, the names of dear ones, etc.). There are known attacks of various types that drastically reduce the protection provided by this type of scheme.

Other known methods include the steps of storing entire portions of software on the token, but in this way there is an impact on software's speed of execution, linked to the fact that it is necessary to access these portions, compile them and then execute them. In consequence, the check is usually carried out only in the software's start-up phase and not during all of the time period in which the software is used.

The object of the present invention is to provide a data protection system and method able to overcome the drawbacks of the prior art.

In greater detail, the authentication system is based on a token accessible by means of a PIN, and is configured to authorize or deny access to information to be protected, the token having a memory that stores: an array containing numerical and/or alphanumerical and/or alphabetic information items (I₁-I_N), and random numerical and/or alphanumerical and/or alphabetic data interposed between immediately consecutive information items in a reading order of the array; and an encrypted seal scheme vector (or seal scheme array), containing information regarding the arrangement of the information items in the array and information regarding the size (typically in bytes) of each random data item, so as to enable access to each of the information items in their respective positions in the array. The authentication system is configured to: subject access to the token to entering a correct PIN; decrypt the seal scheme vector; acquire the arrangement information of the information items and the information on the space occupied by each random data item from the seal scheme vector; check correspondence between the acquired arrangement information and the effective arrangement of the information items in the array, and between the acquired size information and the effective size of the random data; authorize or deny access to the information to be protected on the basis of a result of the previous check.

Document US2011289576 is directed to a method and apparatus for providing a token for secured access.

According to the present invention, a data protection system and method are provided as defined in the appended claims.

For a better understanding of the present invention, some preferred embodiments will now be described, purely by means of non-limitative example and with reference to the attached drawings, where:

• Figure 1 schematically shows a device, or token, for generating authentication keys;
• Figure 2 shows a computer, of a type in itself known, suitable for being operatively connected to the token in Figure 1;
• Figure 3 shows, by means of a flowchart, the steps of a data protection method according to one embodiment of the present invention;
• Figure 4 shows, in graphical form, an array comprising a plurality of data/information items encoded according to steps 10 and 12 of the method in Figure 3;
• Figure 5 shows the array of Figure 4 according to a different form of graphical representation;
• Figure 6 shows, by means of a flowchart, the sub-steps of the method in Figure 3;
• Figure 7 shows, by means of a flowchart, further sub-steps of the method in Figure 3; and
• Figure 8 shows, by means of a flowchart, further steps of the data protection method according to the present invention.

The present description relates to a system and method for protecting data and, in particular, to cypher or encrypt electronic data.

More in detail, the levels of protection provided by the present invention comprise: a physical level, including an authentication device or token, able to generate a pseudorandom authentication code, for example, in a numeric, alphabetic or alphanumeric format; and a software level, including cryptographic steps performed by means of a computer program.

The authentication token, schematically shown in Figure 1 and indicated by reference numeral 1, comprises a microprocessor 2 of known type, able to generate, upon user request, the pseudorandom authentication code according to a certain algorithm, which can take a plurality of factors into account, such as, for example, the serial number of the authentication token 1, a current date value, a time value defined by a clock inside the authentication token 1, or other factors. This value is generated via hardware by a time-based algorithm and can vary between different token manufacturers. Typically, how the pseudorandom authentication code is generated by the token is not disclosed and each token supplier adopts a proprietary algorithm. The method according to the present invention is independent of the way in which this authentication code is generated and so the latter is not described any further herein.

The authentication token 1 also comprises a memory 4, for example a flash-type memory, connected to the microprocessor 2.

The program for generating the authentication code resides in the memory 4 of the authentication token 1, in particular in a private area 4a of the memory 4, to be used for storing sensitive data accessible by means of a keyword, or PIN (Personal Identification Number). A public area 4a of the memory 4 can store certificates of a public nature and non-sensitive information.

The PIN is generated automatically and is not made available to the end user. Upon installation, or when the software is first run, the user is prompted to choose a password of his/her liking that unlocks utilization of the PIN and stores the hash (i.e. the result of an operation that is non-reversible or reversible "with difficulty" - in a database).

The private area 4a of the memory 4 can be of the hidden type (the user is not aware of its presence), or visible to the user, but with access protected by a keyword or password. This password can be modified, subject to having the seed record generated by the supplier and the authorization of the token administrator (known as the Security Officer) that is protected by another password, typically modifiable and not overwritable, which is set to a predetermined value and saved in a secure location.

The private area 4a of the memory 4 can be used to store a plurality of data and/or information, used during the steps of the cryptographic method, as better illustrated hereinafter.

In one embodiment, the authentication token 1 can be connected to a computer 6 (Figure 2), so that information can be exchanged between the authentication token 1 and the computer 6. For example, the authentication token 1 could be equipped with a USB port for connection to the computer 6.

Hereinafter, reference is made to a plain text that it is wished to encode in order to prevent an unauthorized user accessing the information contained in the plain text.

With reference to Figure 3, a flowchart is shown that includes steps of the encryption method according to one embodiment of the present invention.

According to the present invention, a "seal" is generated that includes a plurality of data and information, chosen on a discretional basis, which is encrypted according to a random scheme. The decryption and reading of this seal is a requisite for authorizing or denying one or more further activities.

According to a preferred application, the decryption and reading of this seal is a requisite for authorizing the execution of a software program, or obtaining a license to use this software program. The encryption steps according to the method of the present invention are not therefore aimed at encrypting the software program to be executed, but to generating an encrypted seal, the reading of which is necessary for obtaining authorization to execute this software program.

The method according to Figure 3 comprises initialization steps (steps 10 and 12) during which the seal is generated from the plain text and is written in the private area 4a of the memory 4, and encryption steps (steps 14 and 16) during which a cryptographic method of known type (for example, a known method chosen from RSA, AES, DES, or others) is used to encrypt the seal generated in steps 10 and 12 in order to guarantee a further level of protection.

The seal is generated on the basis of the data and/or information that constitute the plain text, arranging this data in an array as better described hereinafter. In addition, the seal has a size, in bits, such as to be compatible with the encryption algorithm used in steps 14 an 16 (for example, a length of 128 bits, or 196 bits, or 256 bits, or some other size) and leave sufficient "free" spaces to fill with random data.

The plain text can comprise a plurality of data and/or information of a fixed type (the value or information content of which does not vary with time) and/or a variable type (the value or information content of which is not the same if considered at different moments in time). Fixed data includes, for example, predetermined identification codes in the form of numbers, letters and/or alphanumerics or some other format, such as, for example, one or more of: trade name or code name of the software, name of the software manufacturer, progressive license number for the current user of the software, customer name and/or identification code, initialization date, token serial number, software modules to enable (in the case of software composed of multiple modules) and associated license password/codes, and other optional data.

Variable data includes, for example, values for the current date, current time, or other data that can take a variable value on the basis of predetermined or unplanned conditions.

With regard to variable fields, the correctness of their content is checked "by appointment". In other words, the software component checks that, given a time value written at time t₀, after "x" seconds have passed, the time value is time t_o+x (plus or minus a tolerance of the order of hundreds of milliseconds). The value t₀ is the moment when the date/time value written and "x" is a variable offset determined instantaneously and randomly by the software according to a freely chosen uniform distribution, the limits of which are specified between a minimum and a maximum, usually of the order of tens of seconds/minutes.

Hereinafter in the description, each data item, fixed or variable, is considered as a single "item" and indicated by references I₁, I₂, I₃, ..., I_N. A plurality N of items I₁-I_N of this type form a set (arranged in the form of a vector I) comprising a number N of items I₁-I_N: $$\mathrm{I}=\left\{{\mathrm{I}}_1,I_2,I_3,I_4,I_5,\dots ,I_{\mathrm{N}}\right\}$$

As mentioned, one application of the present invention is to authorize or deny execution of a software program, or to obtain a license to use this software program. In this case, a copy of the vector I is saved in the software program in a manner that is not accessible to a user of the software program.

When considered in their digital format (as a sequence of bits with logical values "1" or "0"), each item I₁-I_N of the set of N items I₁-I_N has its own length L(i), measurable as a number of bits and/or bytes, where 1 byte is a group of 8 bits. In this case, "i" takes values comprised between 1 and N, and therefore L(1) is the length of the first item of vector I, L(2) is the length of the second item of vector I, and so on, where L(N) is the length of the N'th (last) item of vector I.

Furthermore, each item I₁-I_N has its own value J(i), or J₁-J_N. For example, if the item I₁-I_N is a date, the value J(i) of this item is an actual date, for example in the day/month/year format. If the item I₁-I_N is the name of the software manufacturer, the value J(i) of that item is the name of the manufacturer, and so on. Therefore, for each item I₁-I_N, the value J(i) has, indifferently, a numeric, alphanumeric or alphabetic value.

The above-mentioned items I₁-I_N are arranged in an array 100 having a plurality M of fields defined by the intersection of a row x (with x=x₁, x₂, ... , x_R) with a respective column y (with y=y₁, y₂, ..., y_C). The array 100 is represented graphically in Figure 4. The N items I₁-I_N each occupy a respective field (x, y) of the array 100. The array 100 has a number of fields equal to M=x_R·y_C and greater than N. In particular, the value of M is chosen so that the items I₁-I_N are stored in respective fields (x, y) that are not directly consecutive to each other and so as to support the chosen encryption algorithm. In other words, the field immediately following, along the rows of the array 100, a field containing an item I₁-I_N, does not contain another of the items I₁-I_N.

When all the items I₁-I_N have been written in the array 100, the fields (x, y) of the array 100 left free are filled by randomly generated bytes, by means of a random number generator of known type. In this way, each item I₁-I_N is separated from the other items I₁-I_N by one or more random bytes.

In Figure 4, the random bytes are indicated by RB_x,y, where subscripts x,y take the value of the corresponding row x and the corresponding column y of the field in which the random bytes are written. For example, one or more random bytes RB_1,1 are stored in the field identified by the first row x=x₁ and the first column y=y₁. Similarly, one or more random bytes RB_2,2 are stored in the field identified by the second row x=x₂ and the second column y=y₂.

Always with reference to Figure 4, item I₁ occupies the field corresponding to the first row x=x₁ and the second column y=y₂, and is indicated as I¹_1,2; similarly, item I₂ occupies the field corresponding to the second row x=x₂ and the first column y=y₁, and is indicated as I²_2,1; and so on.

As can be noted in Figure 3, the random bytes RB_x,y are inserted both between the items I₁-I_N (when considered in sequence along the rows x₁-x_R) and before the first item I₁ of the array 100, as well as after the last item I_N of the array 100. In other words, according to this embodiment, both the first field of the array 100 and the last field of the array 100 contain random bytes RB_x,y.

It is clear that the length in bytes of each item I₁-I_N and of each sequence of random bytes RB_x,y can have any length, in bytes. Fields x,y of the array 100 that are different from each other can contain a different number of bytes. This situation is shown by way of a graphical example in Figure 5, where random bytes are graphically shown by fields filled with hatched lines and items I₁-I_N by blank fields.

According to one embodiment, each field (x, y) of the array 100 is unambiguously identified by an identification number less than or equal to M, and vice versa. The association of an identification number to each field of the array 100 is implemented, for example, by progressively traversing the array 100 along the rows x, so that the number "1" corresponds to field (x=1, y=1), number "2" corresponds to field (x=1, y=2), and so on, so that number "M" corresponds to field (x=x_R, y=y_C). It is clear that any other association is also possible.

Returning to Figure 3, step 10 comprises defining a set V_I of N integer values identifying the reciprocal order in which the items I₁-I_N must be inserted in the array 100. The order is, for example, considered as starting from the first field of the array 100 and traversing the rows x. However, any other order can be chosen.

As shown in Figure 4, each item I₁-I_N is separated from the others items I₁-I_N by a field (x, y) of the array 100 containing a sequence of random bytes RB_x,y. To this end, in step 10 of Figure 3, a set V_RB of N+1 random numbers is generated in which each one of these numbers identifies the length, in bytes, of a respective sequence RB_x,y which must be inserted in the array 100 as a separator of the items I₁-I_N.

Therefore, the (vector) set V_I identifies the order in which to insert the items I₁-I_N in the array 100 and the (vector) set V_RB identifies the length of each block RB_x,y that must be inserted between one item I₁-I_N-1 and the next item I₂-I_N (and also at the beginning and end of the array 100). The vectors V_I and V_RB are saved in a seal scheme vector SS having the following form (the numeric values indicated in the seal scheme vector SS are illustrative of a possible embodiment and are not limitative): $$\mathrm{SS}=\left\{{\mathrm{V}}_{\mathrm{I}};V_{\mathrm{RB}}\right\}=\left\{\mathrm{1,2,3,4,}\dots ;\mathrm{}\mathrm{23,4,46,5,11,}\dots \right\}.$$

In this way, in the absence of the seal scheme vector SS, the simple reading of the array 100 does not allow easily recovering the information (i.e. it is not possible to immediately trace the plain text, or rather the items I₁-I_N) contained in the array 100.

Vector V_I comprises N fields (and therefore N values in the case where each field contains a single value), and vector V_RB comprises N+1 fields (and consequently N+1 values in the case where each field contains just one value); the seal scheme vector therefore comprises 2N+1 values, where the value of SS(N+1) is the number of bytes that precede the first item I₁ in the array 100; the value of SS(N+2) is the number of bytes between the first item I₁ and the second item I₂ in the array 100; the value of SS(N+3) is the number of bytes between the second item I₂ and the third item I₃ in the array 100; and so on. The value of SS(2N+1) is the number of bytes added in the array 100 after the last item I_N.

As previously mentioned, the step of defining the (vector) set V_I is the equivalent of defining N identification numbers of an order in which to reciprocally insert the items I₁-I_N in the array 100. It should be noted that the N values that form vector V_I do not identify respective fields of the array 100, but represent the order of insertion of the items I₁-I_N in the array 100, for example, when considered in sequence along the rows x of the array 100. Similarly to what has been described hereto, vector V_I is, by way of example, of the type: $${\mathrm{V}}_{\mathrm{I}}=\left\{\mathrm{1,}\mathrm{}\mathrm{2,}\mathrm{}\mathrm{3,}\mathrm{}\mathrm{4,}\mathrm{}\dots \right\}.$$

Vector V_I is interpreted in the following manner: insert item I₁ as the first item in the array 100 (I(V_I(1))={I₁}, because V_I(1)={1} and I(1)=I₁); insert item I₂ as the second item in the array 100 (I(V_I(2))={I₂}, because V_I(2)={2} and I(2)=I₂); insert item I₃ as the third item in the array 100 (I(V_I(3))={I₃}, because V_I(3)={3} and I(3)=I₃); insert item I₄ as the fourth item in the array 100 (I(V_I(4))={I₄}, because V_I(4)={4} and I(4)=I₄); and so on for all N items I₁-I_N of vector I.

Always as a qualitative example, the step of defining the set of N+1 random numbers in which each of these numbers identifies the length of a respective block RB_x,y results in the generation of vector V_RB, of the following type: $${\mathrm{V}}_{\mathrm{RB}}=\left\{\mathrm{23,}\mathrm{}\mathrm{4,}\mathrm{}\mathrm{46,}\mathrm{}\mathrm{5,}\mathrm{}\mathrm{11,}\mathrm{}\dots \right\}.$$

Vector V_RB is interpreted in the following manner: insert twenty three random bytes immediately before the first item in the array 100, as defined by vector V_I (in this example, before item I₁); insert four random bytes immediately after the first item in the array 100, as defined by vector V_I (in this example, after item I₁); insert forty six random bytes immediately after the second item in the array 100, as defined by vector V_I (in this example, after item I₂); insert five random bytes immediately after the third item in the array 100, as defined by vector V_I (in this example, after item I₃); insert eleven random bytes immediately after the fourth item in the array 100, as defined by vector V_I (in this example, after item I₄).

The ordering operation of the items I₁-I_N in the array 100 (i.e. generating vector V_I, steps 10 and 12 in Figure 3), and/or the generation of vector V_RB, can be performed any number of times, and enables cyclic updating of the array 100, at regular intervals and/or following special predetermined events. The seal scheme vector SS is also updated in a corresponding manner. In particular, given N items I₁-I_N, these have N! possible permutations before the starting order of the array 100 is obtained again. This means that the ordering of the items I₁-I_N as defined by vector V_I can be changed N! times before returning to the initial order.

In practice, having chosen two first integer numbers "A" and "B", with values between 1 and M, a shuffle function is performed so that the following operations are performed A times:

1. (1) permuting the position of a first value of vector V_I with the position of a second value of vector V_I (the two positions being chosen arbitrarily or by using two different counters); and
2. (2) swapping the position of one or more values of vector V_I with the position of respective one or more values of vector V_I, these positions being after the position in which the first value mentioned in the previous point (1) was originally contained.

The permuting step (1) is performed "B" times on "B" different values, so as to generate a vector V_I that contains values ordered in a manner that is not predeterminable beforehand.

The value of B is chosen randomly and such that B is not a divisor of N!·(N+1)!. In this way, the two numbers B and N!·(N+1)! are mutual primes and, exploiting the fact that they are not perfectly divisible, returning again to the same values after a certain number of repetitions is avoided.

In greater detail, the permuting operation (1) comprises performing a permutation on the numeric values contained in respective positions of vector V_I. As previously described, vector V_I is actually a vector of integer numbers, each of which describes in which position of the array 100 a respective item I₁-I_N (or, rather the value J₁-J_N associated with each respective item I₁-I_N) is inserted. The permuting step therefore comprises the steps of:

• (a) - selecting a first numeric value contained in a respective first field of vector V_I (e.g., V_I(1), containing the number "1");
• (b) - selecting a second numeric value contained in a respective second field, different from the first field, of vector V_I;
• (c) - swapping the first numeric value contained in the first field and the second numeric value contained in the second field with each other.

The permuting step also comprises the optional step of:

• (d)- repeating steps (a)-(c).

Figures 6 and 7 show the above-mentioned steps (a)-(d) in greater detail, by means of flowcharts.

With reference to Figure 6, in step 20, it is assumed to be at the first iteration of the method in Figure 6. Vector V_I is, for example, of the type V_I={1, 2, 3, 4}. Variable p is an index that identifies, in this example, a field of vector V_I and can take values from 1 to (N-1), in this case (N-1)=4. The penultimate position of vector V_I, which is V_I(p)=V_I(N-1)=V_I(3)={3}, is selected.

Then, in step 22, the last field of vector V_I is selected, i.e. V_I(p+1)=V_I(N)=V_I(4)={4}.

Afterwards, in step 24, the values contained in fields V_I(N-1) and V_I(N) are compared to each other. If V_I(N-1) is less than V_I(N), then, in step 26, V_I(N-1) is selected as the first item according to the above-mentioned step (a).

Otherwise, in step 28, value p is decremented by one unit and it is checked, in step 29, if the beginning of vector V_I has been reached. If the current value of p is equal to 0 then vector V_I has been exited and all possible permutations have been made (the method ends). If p is greater than or equal to 1 then, in step 30, V_I(p) is selected; and, in step 32, V_I(p+1) is selected. Processing then returns to step 24 for comparing V_I(p) with V_I(p+1).

Continuing with the previous example, given vector V_I={1, 2, 3, 4}, step 24 leads to step 26, as V_I(N-1)<V_I(N).

Steps 28-32 are not performed and value V_I(N-1)=3 is thus selected as the first numeric value of the above-mentioned step (a).

Processing then passes to the selection of the second numeric value, according to the above-mentioned step (b). With reference to Figure 7, in step 36, the last value of vector V_I (i.e. V_I(N)) is selected in the first iteration of the method in Figure 7. According to the method in Figure 7, index q is used to index vector V_I. Therefore, in step 36, q=N.

Then, in step 38, the value V_I(p) selected in step 26 in Figure 6 is compared with the current value V_I(q). If V_I(q)>V_I(p), then, in step 40, the value contained in field V_I(p) is swapped with the value contained in field V_I(q). In this case, always considering vector V_I={1, 2, 3, 4}, step 40 leads to the generation of vector V_I={1, 2, 4, 3}.

If the comparison in step 38 reveals V_I(q)<V_I(p), then, in step 42, value q is decremented by one unit. In step 43, it is checked whether the current value of q is greater than the current value of p. If true, processing returns to step 38 for a new comparison between V_I(p) and V_I(q); if false, the method ends.

Then, processing passes to the previously mentioned "swapping" step. During this step, the numeric values contained in the N-p positions of vector V_I that follow the p'th position are subdivided into two subgroups of (N-p)/2 values each. The position of the two subgroups of (N-p)/2 values are then swapped. In the case where N-p is an odd number, it is possible to consider N-p+1 fields, or choose a different group.

Returning to vector V_I generated after step 40 in Figure 7, i.e. V_I={1, 2, 4, 3}, the value of the index p is equal to 3 and the swap step has no effect.

By iterating steps 20-32 in Figure 6 and Figure 7 again (second iteration), the new V_I vector that is obtained has the following form: $${\mathrm{V}}_{\mathrm{I}}=\left\{\mathrm{1,}\mathrm{}\mathrm{3,}\mathrm{}\mathrm{4,}\mathrm{}2\right\}.$$

Then, performing the swap step gives: $${\mathrm{V}}_{\mathrm{I}}{}^{\left(2\right)}=\left\{\mathrm{1,}\mathrm{}\mathrm{3,}\mathrm{}\mathrm{2,}\mathrm{}4\right\}.$$

By iterating steps 20-32 in Figure 6 and Figure 7 again (third iteration), the vector V_I that is obtained has the following form: $${\mathrm{V}}_{\mathrm{I}}{}^{\left(3\right)}=\left\{\mathrm{1,}\mathrm{}\mathrm{3,}\mathrm{}\mathrm{4,}\mathrm{}2\right\}.$$

In this case, the swap step does not change the vector V_I={1, 3, 2, 4}.

By iterating steps 20-32 in Figure 6 and Figure 7 again (fourth iteration), the vector V_I that is obtained has the following form: $${\mathrm{V}}_{\mathrm{I}}=\left\{\mathrm{1,}\mathrm{}\mathrm{4,}\mathrm{}\mathrm{3,}\mathrm{}2\right\}.$$

Then, performing the swap step gives: $${\mathrm{V}}_{\mathrm{I}}{}^{\left(4\right)}=\left\{\mathrm{1,}\mathrm{}\mathrm{4,}\mathrm{}\mathrm{2,}\mathrm{}3\right\}.$$

Processing continues in this way, obtaining all the possible permutations of the numeric values contained in vector V_I.

In order to make the order of the items I₁-I_N in the array 100 random at the time of its first generation, it is possible to perform iterations of the steps of the methods in Figures 6-7 and of the swap step "A" times (where A is a random number), so as to generate an order of the numeric values in vector V_I that is not predeterminable beforehand.

According to a further embodiment, each iteration of the steps of the methods in Figures 6-7 is performed "B" times before performing the swap step.

The step of defining the (vector) set V_RB of N+1 random numbers, where each of these numbers identifies the length in bytes of a respective block RB_x,y, satisfies at least one of the following requirements:

1. (1) at least one block RB_x,y (of any size/length in bytes) is inserted before the first item I₁-I_N in the array 100;
2. (2) at least one block RB_x,y (of any size/length in bytes) is inserted between two consecutive items I₁-I_N along the rows x of the array 100, so as to separate the two consecutive items I₁-I_N from each other;
3. (3) at least one block RB_x,y (of any size/length in bytes) is inserted after the last item I₁-I_N in the array 100;
4. (4) the probability distribution of the lengths in bytes of each block RB_x,y is uniform; and
5. (5) the sum S_RB of the lengths, in bytes, of all the blocks RB_x,y is such that the array 100 is completely filled (i.e. ∑_iL(i)+S_RB=M).
6. (6)

According to one embodiment of the present invention, the length of a respective block RB_x,y, satisfies all of the previous requirements (1)-(5).

Returning to the flowchart in Figure 3, in step 14, the array 100 generated according to steps 10 and 12 is encrypted using a key k_s generated randomly by the authentication token 1 in Figure 1. Depending on the encryption algorithm chosen, key k_s can be generated according to a standard method. In the case described by way of example, it is the token that handles its generation. Thus, the encryption step using key k_s invokes this standard routine (and therefore, in itself, not the subject of the present invention). For example, in the AES cryptographic method, the key is as long as the minimum block of encryptable data (typically 32 bytes) and is generated from an initialization vector that the user can choose or have generated in a random manner.

In addition, the array 100 is encrypted using a cryptographic method of known type, for example an asymmetric encryption algorithm, such as RSA, or a block-based encryption algorithm such as AES (Advanced Encryption Standard), or the DES (Data Encryption Standard) algorithm, or any other encryption algorithm known in the literature. The above-mentioned encryption algorithms RSA, AES and DES are known in the literature and the associated encryption steps are not specified here for brevity. The generation of the keys k_s is also performed, as stated, by methods known in the literature.

An encrypted array (or, in general, an encrypted text object), hereinafter indicated as "C_data", is obtained as a result of step 14 in Figure 3.

Both the key k_s and the encrypted text object C_data are stored in the private area 4a of the memory 4 of the authentication token 1.

Similarly to that described with reference to the encryption of the array 100, the seal scheme vector SS, which represents the read key for the array 100, is also encrypted (step 14) by means of a respective key k_v generated randomly by the authentication token 1 in Figure 1. The generation of key k_v takes place in a similar manner to the generation of key k_s, by means of a freely chosen known method. For example, referring again to the AES algorithm, this key k_v is generated by an initialization vector that is different from the initialization vector used for key k_s.

Vector SS is also encrypted with known algorithms, similarly to the array 100. For example, the AES encryption algorithm could be used. Nevertheless, it is clear that other algorithms can be used.

Both key k_v and the encrypted seal scheme vector SS (hereinafter also indicated as "C_scheme") are saved in the private area 4a of the memory 4 of the authentication token 1. The encrypted text object C_data constitutes a "seal" that holds information (items I₁-I_N) the knowledge of which is a necessary condition to enable access to further data and/or information that it is wished to protect. For example, continuing with reference to the preferred application of the present invention, the decryption and the reading of this seal is requisite for authorizing the execution of a software program, or obtaining a license to use this software program. As previously stated, a copy of vector I that contains items I₁-I_N is saved in the software program in a manner that is not accessible to a user of the software program. Verification of correctly reading the array 100 (using the seal scheme SS) is performed by comparing the items I₁-I_N read in the array 100 with the items I₁-I_N of vector I saved in the software program.

For authorizing access to this further data/information/software program, the steps of a method according to Figure 8 are performed.

First of all, in step 50, a check is made to establish whether the authentication token 1 used corresponds to the expected authentication token 1.

This check can be performed by checking that the model and/or serial number of the authentication token 1 in use are consistent and compatible with those expected.

In the case where the check in step 50 provides a positive result, then, in step 52, access is made to the private area 4a of the memory 4 to verify if the seal is integral or has been altered. Access to the private area 4a of the memory 4 takes place, as stated, with a PIN that is generated in the licence initialization step and consists of a set of printable characters (alphanumeric plus symbols) of sufficient length (for example, 127 characters) to render any attempted attacks ineffective or particularly difficult. In fact, in the event of trying to guess this PIN by successive attempts, if a predetermined number of failed attempts is exceeded, the token is locked in an irrecoverable manner, forcing the software manufacturer to be contacted, who is the only one that can unlock it.

According to one embodiment, the following steps are performed: (a) when the user attempts to start the software to be protected, the latter prompts the user to insert the token and a personal password; (b) the user provides his/her personal password as requested in previous point (a); (c) the software to be protected checks, with its own method that is not part of the present invention, that the user's personal password is correct (e.g., in the case where a hash function is used, it computes the hash and checks it with a previously stored hash, for example in a file, in a database, on a server, etc.); (d) if the check in the previous point (c) is passed, the software to be protected uses the PIN corresponding to the user's license to open communications with the token and perform the relevant checks according to the steps of the method of the present invention.

Then, in step 54, using keys k_s and k_v, the encrypted text object C_data and the encrypted seal scheme SS C_scheme are decrypted, in a known manner using the encryption algorithm used according to steps 14 and 16 in Figure 3. The consistency is then checked between the positions of the items I₁-I_N in the array 100 as identified by the seal scheme SS and the effective positions of the items I₁-I_N in the array 100.

In a first check, by reading the positions specified in the seal scheme SS, it is verified that the content read from the array 100 is consistent with the items I₁-I_N of vector I contained in the software program that it is wished to protect, for example by checking one or more of the following parameters: (i) field cardinality; (ii) field length; (iii) field type (number, string, date, etc.); (iv) content of fixed fields (e.g., name of software manufacturer); (v) consistency between software and the software name specified in the seal; (vi) consistency between information regarding the hardware that contains the seal (for example, the serial number of the token chip), obtained by interrogating the hardware, and the equivalent data inscribed in the seal. In particular, this last point (vi) implies that in the case where an ill-intentioned person succeeds in totally copying (without destroying the chip at the same time) the content of private memory area 4a into another token of the same model, access to the data would still be denied, because it is tied to that particular piece of hardware (token).

In this step, the method also provides for reading the variable information. By way of non-limitative example, this information comprises one or more of the following: license number; name of client; date/time of generation of current seal; modules activated by the license (if applicable).

In the case where access to the data/information to be protected is denied, step 55, the method in Figure 8 ends.

In the case where access to the data/information to be protected is allowed, then, in step 56, a new seal scheme vector SS is randomly generated and, on the basis of the new seal scheme vector SS thus generated, a new array similar to array 100 is generated. According to one embodiment, the new array contains the same items I₁-I_N of the array 100, arranged according to a new order and separated by blocks RB_x,y having different numbers of bytes. Alternatively, according to a different embodiment, the new array contains items I₁^(a)-I_N'^(a) different from items I₁-I_N of array 100; the number N' of items I₁^(a)-I_N'^(a) can also be different from the number N of items I₁-I_N. According to a further embodiment, the new array contains items I₁^(a)-I_N'^(a) that are partly common to items I₁-I_N of the array 100 (for example, the fixed fields are the same) and partly different (for example, the variable fields could be different). In this way, a new seal is generated.

The data fields used for generating the new seal are the same ones of the original seal as already described hereinbefore. The date/time data is provided by the local computing device on which the software is running during the generation of the new seal. The scheme is made to evolve from the current permutation in which it finds itself. The reserved spaces for the random bytes (RB_xy) are reallocated in a random manner (using the same previously described criterion).

Every time the seal is made to evolve according to step 56 in Figure 8, the encryption steps according to steps 14-16 in Figure 3, not described here for brevity, are also performed.

Then, after a certain predetermined or random period of time, consistency between current seal scheme SS and the current array is checked again, as described with reference to step 54 .

Steps 54-56 continue to be iterated until step 54 provides a result considered to be acceptable, i.e. such that the arrangement of the items I₁^(a)-I_N'^(a) in the current array is consistent with the arrangement as identified by the current seal scheme SS, and the sizes in bytes of the blocks RB_x,y are consistent with the sizes specified by the current seal scheme SS.

When step 54 provides a positive result as output (correspondence between vector SS and array 100), then the token used is considered valid and authorization is obtained to access the subsequent data and/or information (e.g., authorization to execute a software program, or obtain a license to use this software program).

From this moment on and for the duration of the current usage session of the software, it is only checked that the token hardware continues to respond with its own identity, which has been previously confirmed.

Nevertheless, the previously described in-depth checking can be invoked again by the software on the basis of the event of the user requesting to use functions that are considered critical for using the software (or those that the manufacturer considers of greater value).

In the event of the token being disconnected, this situation is detected and reinsertion of the same token is required (by "same", the one with that particular serial number is intended), with the reinitialization of the entire checkout procedure once it is inserted.

Otherwise, if the token is not inserted or has a different serial number, the software aborts its execution (which means that once a session is started with one token, it is not possible to continue with another token, even if this other token contains a valid license).

The advantages of the system and method according to the present invention are evident.

For the choice of ordering in the vector containing the information items I₁-I_N, the (mathematically provable) fact that one and only one way exists of traversing all of the possible permutations of the N items by swapping the positions of a pair at a time, according to the concept of so-called lexical permutations (obtainable according to the Teinhaus-Johnson-Trotter algorithm), is exploited. This method is computationally efficient and forms a Hamiltonian path, i.e. visits all of the vertices of a hypothetical permutahedron, passing through each vertex only once. In this case, the permutahedron is a hypothetical solid in which each vertex contains a possible combination of the information items I₁-I_N different from the other combinations present on the remaining vertices.

The way of traversing this sequence is chosen according to an ad hoc criterion: in fact, by jumping along the sequence by a number of steps chosen so as not to be an integer divisor of the number of items (which is N! (N+1)!), it is guaranteed that the original sequence will be generated in a unique Hamiltonian cycle. Essentially, the sequence becomes an acyclic loop. In this way, a so-called proof of knowledge system is defined, where the token is the holder of the "knowledge" (or, in other terms, of the seal).

The software component is configured to check the soundness of the seal, which demonstrates the veracity of the license (and therefore serves as "proof" or "witness") present on a generic valid token that holds the test data/information.

Thus, taking the above into account, and that:

• the cryptographic system comprises the functions of initialization of the first seal, opening the seal, and validating the seal;
• the seal changes form each time according to a random scheme that does not alter the content of the license, but changes the arrangement with respect to the random material of the rest of the message;
• the completeness criterion is satisfied, i.e. the verification of each initial and subsequent seal is successful, whatever the scheme adopted for the license;
• the perfect soundness criterion is satisfied, as the probability of having the seal accepted is null if the seal is not in the correct form (e.g., if counterfeited between two successive checks), and so each verification of this system is of the non-interactive zero-knowledge proof type and it should be specified that it uses an evolutive commitment scheme. In conclusion, apart from the obvious increase in security, the advantages obtained are:
  1. (i) even though requiring the transmission of data, the security of the license information is not compromised in the exchange of data between the software and the hardware necessary to validate the license (because it is not revealed); and
  2. (ii) at the same time, the possibility of managing multiple floating licenses on multiple machines with a same software is maintained (i.e. the most general situation where there are a multiple machines and multiple users and each user can work on the machine they prefer).

Regarding the use of the PIN, at the first initialization, the code in question is generated automatically so as to be very long (e.g., 128 characters) and is not disclosed to the user, but is used automatically and transparently at low level by the software. In this way, bearing in mind that the token can be configured to automatically lock itself after a predetermined number of failed attempts (for example 3 or 4), the probability of access to the private area of the token by an unauthorized user wanting to access the content is minimized.

Furthermore, according to the present invention, various levels of protection are provided:

1. (1) a physical level, which requires access to the authentication token 1, and is such that the chip is tamper proof, or rather if it attempted to physically extract the chip, it is practically impossible to recover the information saved therein, which becomes lost/destroyed;
2. (2) a private memory area 4a, which is hidden from the user and only accessible via the password set by the manufacturer of the software;
3. (3) a cryptographic level, which enables further protection of the information in the event of the private area 4a of the memory being discovered and/or its access password becoming known;
4. (4) a reciprocal arrangement scheme of the items I₁-I_N such that the simple reading of this scheme or of the array generated from this scheme does not allow the data arranged according to this scheme to be easily recovered;
5. (5) a consistency verification mechanism such that access or decryption attempts that cause even minimum changes to the data stored in the array trigger the invalidation of the seal.
6. (6) the possibility of installing the software on multiple machines and implementing a floating license that only enables the machines on which the token(s) containing the license(s) is/are inserted. For example, installation on 10 machines, but with only 3 tokens.

Finally, it is clear that modifications and variants can be made to the invention described and illustrated herein without leaving the scope of protection of the present invention, as defined in the appended claims.

For example, the private area 4a of the memory 4 could accommodate a plurality of seals generated as previously described, so as to simultaneously manage a plurality of software and users.